Thermal power unit load self-adaptive prediction control method and system

Abstract

This application relates to the field of thermal power generation technology and discloses a method and system for adaptive predictive control of thermal power unit load. The method includes: collecting operating load and environmental parameters, analyzing the temporal pattern of load deviation, and determining the deviation characteristic description; analyzing the correlation between load deviation and equipment status based on the deviation characteristic description, and determining whether it is long-term wear or short-term disturbance; for long-term wear, assessing and quantifying the efficiency impact, predicting the load fluctuation trend and deviation evolution path, and generating preliminary adjustment parameters; dynamically adjusting operating parameters based on the preliminary adjustment parameters until the deviation meets a threshold, and determining the stable operating parameter configuration; for short-term disturbances, monitoring changes in the disturbance amplitude, verifying the system status after recovery, and generating recovery commands; integrating the stable configuration and recovery commands to generate a comprehensive control sequence, verifying its environmental adaptability, and outputting the final adjustment scheme. This application improves the accuracy, adaptability, and operational stability of thermal power unit load control.

Description

Technical Field

[0001] This application relates to the field of thermal power generation technology, and in particular to a method and system for adaptive predictive control of thermal power unit load. Background Technology

[0002] As the core power generation equipment in the power system, the operational stability and efficiency of thermal power units directly affect the security and economy of the power grid. With the large-scale grid connection of new energy sources, the power grid places higher demands on the load regulation capabilities of thermal power units, requiring them to frequently participate in grid peak shaving, leading to increasingly severe load fluctuations. However, in actual operation, unit load fluctuations are often accompanied by varying degrees of operational deviations. If these deviations are not identified and effectively controlled in a timely manner, they will result in decreased combustion efficiency, increased energy consumption, and even accelerated equipment aging.

[0003] Currently, most control methods for load fluctuations in thermal power units employ fixed-parameter proportional-integral-derivative (PID) control or simple logic judgment, lacking in-depth analysis and targeted treatment of the causes of deviations. On the one hand, existing methods struggle to distinguish whether load deviations are caused by long-term wear due to changes in the internal state of the equipment or by temporary disturbances caused by external environmental or grid fluctuations, leading to a mismatch between the control strategy and the deviation type, resulting in poor adjustment effects. On the other hand, most methods focus on post-event correction, failing to fully utilize real-time monitoring data and historical operating experience to predict load fluctuation trends, thus failing to achieve proactive control and often only responding passively after the deviation has widened. Furthermore, existing control methods do not adequately consider the impact of environmental parameters. Environmental factors such as temperature and humidity have a significant impact on unit operating conditions, but traditional methods fail to effectively integrate them into the control model, resulting in poor control adaptability under different seasons or climatic conditions. Simultaneously, performance degradation caused by long-term equipment operation is difficult to compensate for in a timely manner in the control strategy, causing unit efficiency to gradually decline over time.

[0004] Therefore, how to accurately predict load fluctuations in thermal power units, intelligently identify the causes of deviations, and adaptively adjust control parameters to improve unit operating efficiency and stability has become a pressing technical problem to be solved in this field. Summary of the Invention

[0005] To address the aforementioned technical problems, this application provides a method and system for adaptive predictive control of thermal power unit load, which improves the accuracy and operational stability of thermal power unit load control.

[0006] In a first aspect, this application provides a method for adaptive predictive control of thermal power unit load, the method comprising: Step S1: Obtain the operating load data and environmental parameters of the thermal power unit, analyze the temporal variation law of the load deviation, and determine the deviation characteristic description; Step S2: Based on the deviation characteristic description, analyze the correlation between load deviation and equipment operating status, and determine the nature classification result of load deviation. The nature classification result includes long-term wear type and short-term disturbance type. Step S3: For the aforementioned long-term wear type, assess the impact of load deviation on unit efficiency, determine the quantification results of wear impact, and based on the quantification results of wear impact, predict the future load fluctuation trend and load deviation evolution path of the thermal power unit, and determine the preliminary set of adjustment parameters. Step S4: Based on the preliminary adjustment parameter set, dynamically adjust the operating parameters of the thermal power unit, determine whether the adjusted load deviation meets the preset deviation threshold, and determine the stable operating parameter configuration; Step S5: For the aforementioned transient interference type, continuously monitor the changes in interference amplitude, and verify the system operating status after the interference amplitude recovers to the normal range, and generate interference recovery adjustment instructions based on the verification results; Step S6: Integrate the stable operation parameter configuration with the disturbance recovery adjustment command to generate a comprehensive operation control sequence, verify the adaptability of the control sequence under different environmental conditions, and determine the final load adaptive operation adjustment output.

[0007] Secondly, this application provides a thermal power unit load adaptive predictive control system, the system comprising: The data acquisition module is used to acquire the operating load data and environmental parameters of the thermal power unit, analyze the time-series variation pattern of the load deviation, and determine the characteristic description of the deviation. The judgment module is used to analyze the correlation between load deviation and equipment operating status based on the deviation characteristic description, and to judge the nature classification result of the load deviation, which includes long-term wear type and short-term interference type. The prediction module is used to assess the impact of load deviation on unit efficiency for the long-term wear type, determine the wear impact quantification result, predict the future load fluctuation trend and load deviation evolution path of the thermal power unit based on the wear impact quantification result, and determine the preliminary adjustment parameter set. The adjustment module is used to dynamically adjust the operating parameters of the thermal power unit based on the preliminary adjustment parameter set, determine whether the adjusted load deviation meets the preset deviation threshold, and determine the stable operating parameter configuration. The verification module is used to continuously monitor the changes in the interference amplitude for the transient interference type, and verify the system operation status after the interference amplitude recovers to the normal range, and generate interference recovery adjustment instructions based on the verification results. The configuration module is used to integrate the stable operation parameter configuration with the disturbance recovery adjustment command to generate a comprehensive operation control sequence, verify the adaptability of the control sequence under different environmental conditions, and determine the final load adaptive operation adjustment output.

[0008] Compared with the prior art, the beneficial effects of the present invention are at least as follows: This application first collects operating load and environmental parameters in real time, extracts load change sequences and deviation value sequences, and uses a time series analysis model to generate deviation feature descriptions, providing a reliable data foundation for subsequent diagnosis. On this basis, through a two-layer diagnostic mechanism that combines historical waveform comparison and time series feature judgment, it accurately distinguishes between two different types of deviations: long-term equipment wear and short-term external interference. This avoids the problem of ambiguous identification of deviation causes in traditional methods, enabling control strategies to be targeted at different deviation types.

[0009] For long-term wear types, energy balance analysis is used to quantify efficiency losses, and environmental parameters are incorporated for correction, achieving accurate assessment of wear impact. Based on the assessment results, time series analysis and Monte Carlo simulation are used to predict load fluctuation trends and deviation evolution paths. Proportional-integral-derivative (PID) control logic is then used to calculate deviation compensation values, forming a preliminary set of adjustment parameters. This effectively transforms the assessment from impact evaluation to control strategy generation. Simultaneously, parameters are prioritized based on response time, enabling dynamic adjustment and closed-loop iterative convergence of operating parameters, ensuring that load deviations quickly meet preset threshold requirements.

[0010] For transient disturbances, recovery adjustment commands are generated through disturbance amplitude monitoring and stability verification, effectively ensuring the unit's rapid recovery capability after being subjected to external disturbances. Finally, by integrating stable operating parameter configurations and disturbance recovery adjustment commands, a comprehensive operating control sequence is generated, and its adaptability is verified through multi-environment scenario simulations, ensuring that the final control output maintains stable and reliable control performance under different environmental conditions. This application significantly improves the accuracy, adaptability, and operational stability of thermal power unit load control. Attached Figure Description

[0011] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a flowchart of a thermal power unit load adaptive predictive control method in an embodiment of this application; Figure 2This is a schematic diagram of the structure of a thermal power unit load adaptive predictive control system according to an embodiment of this application. Detailed Implementation

[0013] This application provides a method and system for adaptive predictive control of thermal power unit load. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0014] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 An embodiment of this application provides a method for adaptive predictive control of thermal power unit load, comprising: Step S1: Obtain the operating load data and environmental parameters of the thermal power unit, analyze the temporal variation law of the load deviation, and determine the deviation characteristic description.

[0015] Step S1 includes: acquiring the operating load data of the thermal power unit through a real-time data acquisition system, and simultaneously acquiring environmental parameters such as temperature and humidity; calculating the load change sequence based on the operating load data, and calculating the difference between the actual load and the expected load as the deviation value sequence; correcting the deviation value sequence using environmental parameters to eliminate the deviation fluctuation component caused by environmental factors, and obtaining the corrected deviation value sequence; determining the duration and amplitude change trend of the deviation based on the corrected deviation value sequence, and generating a deviation feature description.

[0016] Specifically, the operating load data of the thermal power unit is first acquired through a real-time data acquisition system, and external environmental parameters are collected simultaneously. The real-time data acquisition system typically includes a sensor network deployed at key locations of the unit. These sensors collect operating load data such as the unit's output power at a fixed frequency. At the same time, temperature and humidity values ​​are recorded simultaneously through external weather stations or environmental monitoring devices, and the timestamps of all data are kept consistent, thereby providing a reliable time-series alignment basis for subsequent correlation analysis.

[0017] After acquiring the operating load data, it is necessary to further extract the load change sequence and the deviation value sequence. The load change sequence refers to the power difference sequence of thermal power units at consecutive time points, used to characterize the dynamic fluctuation of the load. For example, at two adjacent sampling times, the output power value of the unit at the later time is subtracted from the output power value at the previous time, and the difference is the load change at that time point. Arranging the load changes at all time points in chronological order constitutes the load change sequence. The deviation value sequence refers to the difference sequence between the actual load and the expected load, used to characterize the degree of deviation between the actual operating state and the expected state of the unit. The expected load can be determined based on the unit's historical operating mode or dispatch instructions. For example, in a power grid dispatch scenario, the dispatch center issues a load instruction to the thermal power units based on the load demand of the entire network, and the instruction value is the expected load at the current time. In the absence of dispatch instructions, the historical operating data of the unit under the same operating conditions can be fitted, and the historical load average value within the same environmental conditions and load range can be used as the expected load. Arranging the deviation values ​​at all time points in chronological order constitutes the initial deviation value sequence.

[0018] Since environmental factors such as temperature and humidity directly affect the unit's combustion efficiency, heat dissipation conditions, and steam parameters, even under completely normal operating conditions, fluctuations in environmental parameters can cause deviations between the actual and expected loads. To avoid interference from environmental factors in the subsequent classification of deviation characteristics, it is necessary to correct the initial deviation value sequence using collected environmental parameters to eliminate the deviation fluctuation component caused by environmental factors. The specific correction method is as follows: First, an environmental parameter-deviation influence model is constructed based on historical operating data. At least six months of historical operating data are collected, covering unit operating conditions under different seasons and climates. Environmental temperature and humidity are used as input variables, and the deviation values ​​caused by environmental factors alone under normal operating conditions are used as output variables. The environmental deviation influence function is obtained by fitting multiple linear regression or neural network methods. During real-time operation, the temperature and humidity values ​​collected at the current moment are input into the environmental deviation influence function to calculate the environmental induced deviation value that should have been generated under the current environmental conditions. Then, the environmental induced deviation value at the corresponding moment is subtracted from each deviation value in the initial deviation value sequence to obtain the corrected deviation value sequence. This corrected deviation value sequence eliminates the fluctuation component caused by environmental factors and can more realistically reflect the load deviation caused by changes in the operating status of the equipment itself, providing a more accurate data basis for subsequent deviation nature classification.

[0019] After obtaining the corrected deviation value sequence, it is necessary to extract the duration and amplitude trend of the deviation to generate a deviation feature description. The duration of the deviation is determined as follows: a deviation threshold is set, which is determined based on a certain proportion of the unit's rated load, for example, 3% of the rated load. That is, when the absolute value of the deviation exceeds this proportion, it is considered a valid deviation. Then, the corrected deviation value sequence is traversed to identify each time period in which the absolute value of the deviation continuously exceeds the threshold. The time length from the start time to the end time of each time period is the duration of the deviation event. If there are multiple discontinuous deviation time periods in the sequence, the duration of each time period is recorded separately, and the time period with the longest duration is used as the deviation duration in the current deviation feature description. The trend of deviation change is determined as follows: For the identified deviation period, the corrected deviation value sequence within that period is extracted. Linear regression analysis is performed on this sequence, i.e., a straight line is fitted with time as the independent variable and deviation value as the dependent variable. The slope of this line is calculated. If the slope is positive, it indicates that the deviation amplitude is increasing over time, i.e., the deviation is gradually aggravating; if the slope is negative, it indicates that the deviation amplitude is converging over time, i.e., the deviation is gradually decreasing; if the slope is close to zero, it indicates that the deviation amplitude remains basically stable. In the case of multiple deviation periods, the period with the longest duration or the period with the largest deviation amplitude can be selected for trend analysis according to actual needs. Using the deviation duration and amplitude change trend slope obtained above, a deviation feature description is generated. This description summarizes the temporal pattern of the current load deviation in semantic form, such as "the deviation duration is 25 minutes, and the amplitude is increasing" or "the deviation duration is 3 minutes, and the amplitude is rapidly converging." This deviation feature description will serve as the input basis for the deviation nature classification in the subsequent step S2.

[0020] Through the above process, the systematic processing of thermal power unit operating data was achieved, the accurate extraction of load deviation characteristics and the quantitative elimination of environmental impact were completed, providing an accurate and clean data foundation for subsequent deviation nature classification, thereby improving the pertinence of load control and the stability of unit operation.

[0021] Step S2: Based on the deviation characteristic description, analyze the correlation between load deviation and equipment operating status, and determine the nature classification result of load deviation. The nature classification result includes long-term wear type and short-term disturbance type.

[0022] Step S2 includes: comparing the current deviation value sequence with historical deviation waveform data in the historical database, and calculating the waveform similarity using a dynamic time warping algorithm, wherein the historical deviation waveform data is associated with corresponding equipment status records; if the waveform similarity exceeds a preset threshold, then determining the equipment status change corresponding to the current load deviation based on the associated equipment status records, and locating the nature classification result of the current load deviation based on the equipment status change; otherwise, determining whether the deviation duration in the deviation feature description exceeds a preset threshold and whether the amplitude change trend shows a gradual shift, then determining the nature classification result of the load deviation as long-term wear type; if the deviation duration does not exceed the preset threshold, then the nature classification result of the load deviation is short-term interference type.

[0023] Specifically, after completing the deviation characteristic description, to further distinguish the causes of load deviation, it is necessary to analyze the correlation between the deviation and the equipment operating status, thereby determining whether the deviation belongs to the long-term wear type or the short-term disturbance type, providing a basis for the subsequent formulation of differentiated control strategies. First, the current deviation value sequence is compared with historical deviation waveform data in the historical database. The current deviation value sequence records the change process of the deviation value in chronological order, which is the time-series waveform data of the current load deviation. This data contains the complete evolution trajectory of the deviation from its occurrence to the current moment, providing a basis for subsequent comparison with historical waveforms. At the same time, historical deviation waveform data related to the current deviation characteristics are extracted from the historical database. This historical database pre-stores the complete deviation time-series waveform corresponding to each load deviation event that has occurred in the past, and each waveform data is associated with the deviation cause label diagnosed at that time, such as induced draft fan bearing wear, coal feeder coal blockage, power grid frequency fluctuation, etc. These labels provide an attribution basis for subsequent similarity comparison.

[0024] To determine whether the current deviation has the same cause as historical events, the current load deviation value sequence is compared line by line with historical deviation waveform data in the historical database. A dynamic time warping algorithm is used to calculate waveform similarity. Specifically, the current deviation waveform and the historical deviation waveform are represented as two separate numerical sequences, and a distance matrix is ​​constructed. The value at each position in the matrix represents the numerical difference between the corresponding sampling points of the two waveforms. Then, starting from the upper left corner of the matrix, a path is found to reach the lower right corner. The path can move to the right, down, or diagonally. The sum of the values ​​at all positions along the path is the cumulative difference for that path. The dynamic time warping algorithm iterates through all possible paths to find the path with the smallest cumulative difference. This smallest cumulative difference is the similarity index between the two waveforms; the smaller the cumulative difference, the closer the waveforms are. Each historical deviation waveform data in the historical database is associated with a corresponding equipment status record, such as cause tags like "induced draft fan bearing wear," "coal feeder blockage," or "grid frequency fluctuation." After calculating the similarity between the current deviation waveform and each historical waveform using the above algorithm, the historical deviation waveform with the highest similarity exceeding a preset threshold is selected, and its associated equipment status record is used as a possible cause of the current deviation, thereby achieving accurate attribution of the load deviation cause. If the cause label is a change in equipment status such as bearing wear or valve jamming, it is directly determined to be a long-term wear type, without further verification of real-time time-series characteristics; if the cause label is external interference such as power grid fluctuations or sudden environmental changes, it is directly determined to be a short-term interference type. If the waveform similarity does not exceed the waveform similarity threshold, or there is no matching event in the historical database, the real-time feature supplementation path is used for judgment. That is, the judgment is made based on the duration and amplitude change trend of the deviation in the deviation feature description generated in step S1. When the duration of the deviation exceeds the duration threshold and the amplitude change trend shows a gradual shift, the amplitude change trend shows a gradual shift, which means that the deviation value shows a unidirectional and continuous growth trend over time, rather than a violent instantaneous fluctuation or rapid recovery. Specifically, within the continuous monitoring time window, the deviation amplitude continuously increases, and this increase is cumulative and continuous. For example, the slope obtained by linear regression fitting is significantly positive, indicating that the deviation value is gradually expanding rather than stabilizing or converging. This gradual shift characteristic usually reflects the slow degradation of equipment performance, such as wear, scaling, material fatigue and other internal factors caused by long-term operation. These factors will not suddenly appear in a short period of time, but will gradually appear with the accumulation of operating time, which is manifested as a continuous expansion of the deviation. Therefore, when the deviation characteristic description simultaneously meets two conditions—the duration of the deviation exceeding the duration threshold and the amplitude change trend showing a gradual shift—the nature of the load deviation is classified as long-term wear.

[0025] When the duration of the deviation described in the deviation feature description does not exceed the duration threshold, the nature of the load deviation is classified as a transient disturbance. The duration threshold is determined statistically based on the maximum duration of typical transient disturbance events in historical operating data. The specific statistical method is as follows: collect deviation data marked as transient disturbance events from historical operations, calculate the duration of each event, and take the 95th percentile of these durations as the duration threshold. For example, it can be set as an empirical value between 5 and 30 minutes, with the specific value determined according to the unit type and operating characteristics. When the duration of the current deviation is lower than this threshold, it indicates that the deviation has similar temporal characteristics to historical transient disturbance events and does not possess the gradual cumulative characteristics unique to long-term wear. Therefore, when the duration of the deviation does not exceed the duration threshold, it is classified as a transient disturbance. From a physical mechanism perspective, deviations of the long-term wear type originate from the slow degradation of equipment performance, such as bearing wear, scaling, and material fatigue. These factors usually require a long time to accumulate and form an observable deviation, thus the duration of the deviation is often long. On the other hand, deviations of the transient disturbance type originate from instantaneous disturbances in the external environment or power grid, such as power grid frequency fluctuations, sudden changes in ambient temperature, and instantaneous fluctuations in fuel quality. These disturbances usually manifest as sudden, short-term disturbances, and the deviation recovers quickly after the source of the disturbance disappears, thus the duration of the deviation is short. Based on the above judgment results, a classification result of the nature of the load deviation is generated, which is used to generate subsequent disturbance recovery adjustment commands, thereby realizing differentiated control strategies for different deviation types.

[0026] By combining historical deviation waveform comparison with real-time deviation timing characteristic judgment, a two-layer diagnostic mechanism with historical experience as the priority and real-time characteristics as the supplement is constructed. This enables accurate attribution of the causes of load deviations, effectively distinguishes between two different types of deviations: long-term equipment wear and short-term external interference. This provides a reliable basis for the formulation of subsequent differentiated control strategies and improves the stability and targeting of unit operation.

[0027] Step S3: For long-term wear types, assess the impact of load deviation on unit efficiency, determine the quantitative results of wear impact, predict the future load fluctuation trend and load deviation evolution path of thermal power units based on the quantitative results of wear impact, and determine the preliminary set of adjustment parameters.

[0028] In step S3, assessing the impact of load deviation on unit efficiency and determining the quantitative results of wear impact includes: for the long-term wear type in the nature classification results, obtaining historical equipment operating status data, combining the operating status data, using energy consumption balance analysis method to calculate the impact index of load deviation on unit efficiency; calculating the difference between the impact index and the standard efficiency benchmark, and correcting the difference in combination with environmental parameters, assessing the specific degree of impact of deviation on the overall operating efficiency of the unit, and generating the quantitative results of wear impact.

[0029] Specifically, after classifying the nature of the deviations, for load deviations identified as long-term wear types, it is necessary to further assess their actual impact on unit efficiency and quantify the impact into specific values ​​to provide a data foundation for subsequent load fluctuation trend prediction and adjustment parameter calculation. First, historical equipment operating status data related to this long-term wear type is extracted from the unit's historical database, including key parameters such as boiler temperature, steam pressure, and fuel consumption. This data reflects the performance evolution trajectory of the unit during long-term operation, providing a basis for subsequent energy consumption analysis. Based on this, the energy balance analysis method is used to calculate the impact index of load deviations on unit efficiency. Energy balance analysis is a method to assess energy loss by comparing the balance relationship between input energy and output energy. In practice, firstly, the total input energy is calculated, i.e., fuel consumption multiplied by its lower heating value; then, the output energy is calculated, i.e., the unit's net power generation; the energy loss is obtained by subtracting the output energy from the input energy, and then the energy loss is divided by the total input energy to obtain the efficiency loss rate. This efficiency loss rate is the impact index of load deviations on unit efficiency, and this index directly reflects the degree of energy loss caused by long-term wear.

[0030] To more comprehensively assess the impact of deviations on the overall operating efficiency of the unit, it is necessary to calculate the difference between the aforementioned influencing indicators and the standard efficiency benchmark. The standard efficiency benchmark is determined based on the theoretical efficiency under the unit's design operating conditions, specifically the design efficiency value of the unit under rated load and design environmental conditions. This design efficiency value can be found in the unit's factory technical parameters. If the design value is unavailable, the 95th percentile of the efficiency value within 100 hours of continuous operation after the unit's most recent major overhaul is used as the benchmark. This difference reflects the degree of deviation of the unit's efficiency from the ideal state under the current wear condition. Based on this, the calculated difference is corrected in conjunction with current environmental parameters. Since parameters such as ambient temperature and humidity directly affect the unit's heat dissipation conditions, combustion efficiency, and steam parameters, the efficiency loss caused by the same degree of equipment wear varies under different environmental conditions.

[0031] In specific adjustments, the current ambient temperature and humidity values ​​are first obtained. Then, adjustments are made according to a preset environmental parameter-efficiency correction coefficient table. This correction coefficient is constructed by analyzing the correspondence between the efficiency loss rate under different temperature and humidity conditions and the efficiency loss rate under standard operating conditions in historical operating data. Specifically, at least six months of historical operating data are collected. Using the efficiency loss rate under standard operating conditions as a benchmark, the percentage change in efficiency loss rate for every 1 degree Celsius increase or decrease in temperature is used as the temperature correction coefficient, and the percentage change in efficiency loss rate for every 1 percentage point increase or decrease in humidity is used as the humidity correction coefficient. The formula for calculating the correction coefficient is: The temperature and humidity correction factors are recalculated and updated quarterly based on the latest operating data. The difference is multiplied by the correction factor under the current environmental conditions to obtain the corrected efficiency impact. By introducing environmental parameter corrections, the true impact of deviations under actual operating conditions can be reflected more accurately, avoiding evaluation biases caused by ignoring environmental factors.

[0032] Based on the corrected level of efficiency impact, the specific impact of the deviation on the overall operating efficiency of the unit is assessed, for example, by classifying it into different levels such as mild, moderate, and severe. This level of impact is then converted into a numerical quantification result of wear impact. Specifically, the assessed level of impact is mapped to a preset quantification range, for example, mild corresponds to an efficiency loss of 0% to 1%, moderate corresponds to an efficiency loss of 1% to 3%, and severe corresponds to an efficiency loss of more than 3%. This converts the qualitative level assessment into a quantifiable efficiency loss value, which is the quantification result of wear impact. This quantification result directly reflects the actual degree of loss in the efficiency of the thermal power unit caused by long-term wear. It serves as an input parameter for the subsequent load fluctuation trend prediction model, used to correct the efficiency decay factor in the prediction model, so that the load fluctuation trend prediction can fully consider the impact of wear on the dynamic characteristics of the unit, providing accurate data support for the calculation of adjustment parameters.

[0033] Through the above steps, this embodiment achieves a quantitative impact assessment of long-term wear type deviations, transforming abstract unit efficiency losses into quantifiable numerical indicators, and combining them with environmental parameters for correction, thereby improving the accuracy of the assessment and providing a reliable data foundation for subsequent load fluctuation trend prediction and the determination of adjustment parameter sets.

[0034] Among them, the energy consumption balance analysis method is used to calculate the impact index of load deviation on unit efficiency. This includes: obtaining input energy data and output energy data from historical equipment operating status data, calculating energy loss by comparing the balance relationship between input energy and output energy, and dividing the energy loss by the total input energy to obtain the efficiency loss rate, which serves as the impact index of load deviation on unit efficiency.

[0035] Specifically, after classifying the nature of the deviation, for load deviations determined to be of the long-term wear type, it is necessary to further assess their actual impact on the efficiency of thermal power units and quantify the impact into specific values. This provides a data basis for subsequent load fluctuation trend prediction and adjustment parameter calculation. The calculation of the impact index of load deviation on the efficiency of thermal power units using the energy consumption balance analysis method includes the following process: First, input energy data and output energy data are extracted from historical equipment operating status data. Input energy data refers to the heat energy released by the combustion of fuel in thermal power units, which is obtained by multiplying the coal consumption by its lower heating value, reflecting the total energy input during unit operation. Output energy data refers to the actual electrical energy output by thermal power units, which is obtained by statistically analyzing net power generation, reflecting the effective energy produced during unit operation. Subtracting the output energy from the input energy yields the energy loss of thermal power units during operation. This loss includes ineffective energy consumption caused by factors such as equipment wear, decreased combustion efficiency, and steam leakage. Dividing the energy loss by the total input energy yields the efficiency loss rate, which is the indicator of the impact of load deviation on the efficiency of thermal power units. This indicator is presented as a percentage, directly quantifying the degree of efficiency decline caused by long-term wear. For example, in wear scenarios such as ash accumulation on boiler heating surfaces or scaling on turbine blades, a calculated efficiency loss rate of 2.5% indicates that the overall efficiency of the thermal power unit decreased by 2.5 percentage points due to wear. Through the above energy balance analysis, the abstract unit efficiency loss is transformed into a quantifiable numerical indicator, providing a clear data foundation for subsequent comparisons with standard efficiency benchmarks and adjustments to environmental parameters. This enables a quantitative assessment of the impact of long-term wear type deviations in thermal power units.

[0036] The step S3, which involves determining the set of adjustment parameters, includes: analyzing the future load fluctuation trend using time series analysis based on the quantification results of wear impact, and generating a load fluctuation trend curve; extracting key points from the load fluctuation trend curve, and using Monte Carlo simulation in conjunction with environmental parameters to simulate the deviation evolution path under various environmental scenarios, and determining the possible range of deviation changes under each path; calculating targeted deviation compensation values ​​using proportional-integral-derivative control logic based on the possible range of deviation changes, and summarizing the deviation compensation values ​​to generate a preliminary set of adjustment parameters.

[0037] Specifically, after quantifying the wear impact, in order to transform the assessment results into executable control commands, it is necessary to predict the future load fluctuation trend of the thermal power unit and determine a preliminary set of adjustment parameters based on the prediction results. First, based on the wear impact quantification results, time series analysis is used to analyze the future load fluctuation trend of the thermal power unit. Specifically, an autoregressive integral moving average model is constructed, using the wear impact quantification results, historical load data, and environmental parameters as inputs. First, the load data is tested for stationarity using the augmented Dickey-Fowler test to determine whether the sequence is stationary. If it is not stationary, it is made stationary through differencing, and the differencing order d is determined based on the test results. Then, the autoregressive order p and the moving average order q are determined based on the autocorrelation function and partial autocorrelation function. In the model training phase, load data under historical normal operating conditions are used as training samples, and the least squares method is used to estimate the autoregressive coefficients, moving average coefficients, and exogenous variable coefficients of the model to establish a dynamic relationship model of load changes. In the forecasting phase, the real-time load data, environmental parameters, and quantification results of wear impact at the current moment are used as model inputs. Through iterative calculation, the load forecast values ​​for multiple future moments are obtained. These forecast values ​​are connected in chronological order to generate a load fluctuation trend curve over a future period. This curve reflects the possible direction and magnitude of load changes in thermal power units when considering the wear impact, providing a benchmark for subsequent deviation path simulation.

[0038] Based on this, key points are extracted from the generated load fluctuation trend curve, including load peaks, troughs, and inflection points with large rates of change. These key points represent the main characteristics of load fluctuations and are important reference points for simulating the deviation evolution path. Simultaneously, combined with current environmental parameters, the Monte Carlo simulation method is used to simulate the deviation evolution path under various environmental scenarios. Monte Carlo simulation is a method that generates a large number of possible scenarios through random sampling, effectively handling the uncertainty of environmental parameters. In specific implementation, the fluctuation range and probability distribution of parameters such as temperature and humidity are determined based on historical environmental parameter data. The temperature fluctuation range is taken from the maximum and minimum temperatures in historical data, and its probability distribution adopts a normal distribution, with the mean taken as the current temperature and the fluctuation amplitude taken as the typical amplitude of temperature fluctuations in historical data. The humidity fluctuation range is taken from the minimum and maximum humidity values ​​in historical data, and its probability distribution adopts a uniform distribution, meaning that the probability of each value within the fluctuation range is equal, generating multiple sets of random environmental parameter combinations.

[0039] The dynamic response characteristics of thermal power units are described by a mathematical model that reflects the load's response to regulation commands. This model includes three key parameters: the gain coefficient, reflecting the proportional relationship between the load change amplitude and the regulation command; the time constant, reflecting the speed of the load response; and the delay time, reflecting the lag between the issuance of the command and the onset of load change. These model parameters are identified by analyzing the load response curves after the issuance of step regulation commands in historical operating data. During deviation evolution path simulation, key points of the load fluctuation trend curve are used as input excitations. The model parameters are adjusted according to different combinations of environmental parameters; for example, the gain coefficient increases with rising temperature, and the time constant increases with rising humidity. The evolution trajectory of the deviation over time is obtained through simulation calculations. For each parameter combination, a deviation evolution path is simulated based on the key points of the load fluctuation trend curve and the dynamic response characteristics of the thermal power unit. Statistical analysis of all simulated paths is performed to calculate the distribution characteristics of the deviation amplitude under each path, determining the possible range of deviation changes, including the upper and lower limits of the deviation. This range reflects the statistical distribution characteristics of the deviation under different environmental conditions, providing an input interval for subsequent compensation value calculations.

[0040] Based on the possible range of deviation changes, a targeted deviation compensation value is calculated using proportional-integral-derivative (PID) control logic. PID is a classic feedback control algorithm. In this application, the upper and lower limits of the possible deviation range obtained from Monte Carlo simulation are used as the boundaries of the control target interval. The input to the PID controller is the real-time monitored current load deviation value, i.e., the difference between the actual load and the expected load of the unit. The controller calculates the compensation amount according to the current magnitude, historical cumulative amount, and trend of the deviation value, and then sums them to output a deviation compensation value for the main control quantity. During controller parameter tuning, dynamic adjustments are made based on the possible deviation range obtained from Monte Carlo simulation. Specifically, let the deviation range obtained from Monte Carlo simulation be [L, U], where L is the lower limit of the deviation and U is the upper limit of the deviation. When the real-time deviation value is within this range, it indicates that the current deviation has not exceeded the predicted normal fluctuation range, and the controller outputs a small compensation value or maintains the current control state. When the real-time deviation value exceeds the upper limit U or is lower than the lower limit L, it indicates that the current deviation has deviated from the predicted range, and compensation control is required. The proportional coefficient is tuned based on the deviation range width (UL). The wider the range, the more severe the deviation fluctuation. The proportional coefficient is increased accordingly to enhance the rapid response capability. The integral coefficient is tuned based on the predicted duration of the deviation. The longer the predicted duration, the larger the integral coefficient is to be to eliminate steady-state error. The derivative coefficient is tuned based on the slope of the deviation change trend. The larger the absolute value of the slope, the larger the derivative coefficient is to be to suppress overshoot in advance.

[0041] After the controller parameters are tuned, the proportional-integral-derivative controller receives the current load deviation value e(t) in real time and calculates the compensation for the three components: proportional component. Where Kp is the proportional coefficient, which responds instantly according to the magnitude of the current deviation; integral component ,in The integral component compensates for the historical accumulation of the deviation, eliminating steady-state error; the differential component... ,in The differential coefficient is used to predictively compensate for the deviation based on its changing trend, suppressing overshoot. The sum of the three components yields the deviation compensation value of the main control variable. Because load regulation of thermal power units involves the coordinated actions of multiple actuators, the compensation value of a single main control variable needs to be decomposed into the adjustment amounts of multiple specific operating parameters to form a complete set of adjustment parameters. Therefore, after obtaining the deviation compensation value C of the main control variable, based on the current operating conditions of the thermal power unit and the coordinated control logic, this main control compensation value is mapped to the parameter adjustment amounts of each subsystem. The mapping process is based on a parameter correlation model established from the unit's historical operating data, which reflects the correspondence between the main control variable and the parameters of each actuator. For example, assuming the unit is currently operating at 60% of its rated load, the main control variable deviation compensation value C = +4.2%, indicating that the load needs to be increased by 4.2%. The coordinated control system calculates based on real-time status parameters such as the boiler's thermal storage status, feedwater flow rate, and fuel quantity, according to preset ratios: First, based on the boiler combustion efficiency curve, it is determined that an additional 0.8 tons per hour of fuel is required for every 1% increase in load, therefore the fuel quantity adjustment is 4.2 × 0.8 = 3.36 tons per hour; then, based on the ratio of fuel quantity to feedwater flow rate, an additional 1 ton per hour of fuel quantity requires a simultaneous increase of 0.6 tons per hour of feedwater flow rate, therefore the feedwater flow rate adjustment is 3.36 × 0.6 ≈ 2.0 tons per hour; finally, based on the air-fuel ratio curve, an additional 1 ton per hour of fuel quantity requires an increase of 400 cubic meters per hour of air supply, therefore the air supply adjustment is 3.36 × 400 = 1344 cubic meters per hour. When the main control compensation value is negative, indicating a need to reduce the load, the parameters are reduced accordingly. The fuel quantity adjustment of 3.36 tons per hour, water supply flow adjustment of 2.0 tons per hour, and air supply adjustment of 1344 cubic meters per hour obtained from the decomposition are summarized according to parameter type to form a preliminary set of adjustment parameters containing multiple adjustment commands. This set serves as the basis for the execution of dynamic adjustments in the subsequent step S4.

[0042] Through the above steps, the prediction of future load fluctuation trends of thermal power units and the simulation of deviation evolution paths were realized. Based on the prediction results and combined with environmental uncertainties, targeted deviation compensation values ​​were calculated. Through normalization processing, the reasonable superposition of control components with different dimensions was ensured, forming a preliminary set of adjustment parameters, which provided a parameter basis for subsequent dynamic adjustments, and realized a complete closed loop from impact assessment to control strategy generation.

[0043] Step S4: Based on the preliminary set of adjustment parameters, dynamically adjust the operating parameters of the thermal power unit, determine whether the adjusted load deviation meets the preset deviation threshold, and determine the stable operating parameter configuration.

[0044] Step S4 includes: extracting the deviation compensation range and corresponding response time of each parameter from the preliminary adjustment parameter set; prioritizing the adjustment parameters according to the response time to generate an adjustment priority list; dynamically adjusting the operating parameters with higher deviation severity according to the adjustment priority list; continuously monitoring the change of load deviation value through a real-time data acquisition system after adjustment; if the monitored load deviation value is lower than a preset threshold, the current operating parameter configuration is determined as a stable operating parameter configuration; if the monitored load deviation value is not lower than the preset threshold, historical equipment operating status data is reacquired, the wear impact quantification result is updated, and the adjustment parameters are recalculated, and iterative adjustments are performed until the load deviation value is lower than the preset threshold.

[0045] Specifically, after generating the initial set of adjustment parameters, to achieve dynamic adjustment and deviation closed-loop convergence of the thermal power unit's operating parameters, it is necessary to prioritize each parameter based on its response characteristics and ensure that the load deviation ultimately meets the preset threshold requirements through iterative adjustments. First, the deviation compensation range and corresponding response time of each parameter are extracted from the generated initial set of adjustment parameters. Response time refers to the time required for a certain operating parameter to adjust from its current value to its target value and stabilize; it reflects how quickly the parameter responds to control commands. Response time can be obtained by fitting historical operating data. Specifically, a first-order inertia plus delay model is used to model the parameter adjustment process. The model is as follows: ,in, This represents the change in the parameter at time t relative to its initial value. The steady-state change amplitude of the parameter adjustment. To delay time, This is a time constant. After the adjustment command is issued, the parameters do not change immediately, but rather after a delay. The response only begins after that; subsequently, the parameters change exponentially, with the rate of change decreasing from the time constant. Decide, The smaller the value, the faster the response. The larger the value, the slower the response. The response time is typically defined as the time required for the parameter to reach 95% of its steady-state amplitude. During the fitting process, the change curves of various parameters after the issuance of step adjustment commands were collected from historical operating data, and the least squares method was used to fit the model parameters. , , Estimate the parameters to minimize the sum of squared errors between the model-fitted curve and the actual parameter change curve. Analyze the fitting results under different operating conditions to establish statistical laws governing parameter response times. For example, the response time for fuel quantity adjustment is typically 1 to 2 minutes, for water flow rate adjustment it is 0.5 to 1 minute, and for air supply volume adjustment it is 1.5 to 3 minutes.

[0046] Parameters are prioritized based on their response time, with shorter response times having higher priority and requiring adjustment first. For example, water flow rate has the shortest response time and the highest priority; fuel quantity is next; and air supply volume has the longest response time and the lowest priority. Based on the adjustment priority list, parameters with higher deviation severity are dynamically adjusted first. Deviation severity is quantified by the magnitude of the current load deviation. For instance, when the deviation exceeds a preset severity threshold, fast-response parameters in the priority list are prioritized for adjustment, such as adjusting water flow rate to quickly respond to load changes.

[0047] After adjustment, the load deviation value is continuously monitored through a real-time data acquisition system to calculate the trend of the deviation value and evaluate the adjustment effect. If the monitored load deviation value is lower than the preset threshold, it indicates that the current adjustment has restored the system to a stable operating state. In this case, the current operating parameter configuration is determined as the stable operating parameter configuration. For example, after adjusting the fuel supply rate, if the monitored load deviation value gradually decreases from the initial 8% to 3%, which is lower than the preset threshold of 5%, then the combination of parameters such as the current fuel supply rate, boiler main control command, and feedwater flow rate is determined as the stable operating parameter configuration for the formulation and execution of subsequent control strategies. If the monitored load deviation value is not lower than the preset threshold, it indicates that the current adjustment has not achieved the expected effect. It is necessary to reacquire historical equipment operating status data, update the wear impact quantification results, and recalculate the adjustment parameters based on the updated quantification results for iterative adjustment. This iterative process continues until the load deviation value is lower than the preset threshold, ensuring that the system finally enters a stable state.

[0048] Through the above steps, dynamic adjustment and closed-loop convergence of the operating parameters of thermal power units are realized. The parameters are prioritized according to their response characteristics, and fast-response parameters are adjusted first to quickly suppress deviations. Through iterative adjustment, it is ensured that the load deviation eventually meets the preset threshold requirements, thus forming a stable and reliable operating parameter configuration.

[0049] Step S5: For transient interference types, continuously monitor changes in the interference amplitude, and verify the system operating status after the interference amplitude returns to the normal range. Generate interference recovery adjustment instructions based on the verification results.

[0050] Step S5 includes: for the transient interference type in the nature classification results, calculating the current interference amplitude through real-time load fluctuation monitoring and analyzing the changing trend of the interference amplitude; determining whether the interference amplitude has recovered to the preset normal range; if it has recovered, using time series analysis to perform stability verification on the recovered load data and verify the system operating status; and generating interference recovery adjustment instructions based on the verification results of the operating status.

[0051] Specifically, for identified transient interference types, it is necessary to further monitor the interference's dissipation process and verify system stability after the interference disappears to ensure that the unit can reliably and smoothly recover from the interference state to normal operation. First, for transient interference types in the deviation nature classification results, the current interference amplitude is calculated through real-time load fluctuation monitoring, and the trend of the interference amplitude is analyzed. The interference amplitude refers to the difference sequence between the actual load and the normal load, reflecting the current intensity of the deviation. By comparing the difference changes at adjacent time points, it is identified whether the interference amplitude is increasing, decreasing, or stabilizing, thereby determining whether the interference is dissipating. Based on this, it is determined whether the interference amplitude has recovered to a preset normal range. The normal range refers to a deviation value less than a certain percentage of the unit's rated load, for example, preset to 5% of the rated load. When the interference amplitude gradually decreases from the initial peak to this range, it indicates that the interference source has basically disappeared, and the system has entered the recovery phase. To avoid misjudgment due to instantaneous fluctuations, a recovery confirmation condition is set: the interference amplitude must be below the preset normal range threshold for three consecutive sampling periods (each sampling period is one minute) before it can be determined that the interference has truly recovered. If the threshold is exceeded in any of the three consecutive sampling periods, the counter is reset to zero and counting starts again.

[0052] If the disturbance amplitude has returned to the normal range, time series analysis is used to verify the stability of the recovered load data to confirm whether the system's operating state is truly stable. The stability verification employs an autoregressive moving average model. The model parameters are pre-trained offline using at least 72 hours of load data under historical normal operating conditions. The autoregressive order and moving average order of the model are determined using the autocorrelation function and partial autocorrelation function, enabling the model to accurately describe the dynamic characteristics of the load under stable operating conditions. Once the model training is complete, it is used permanently without requiring online updates. Specifically, the load sequence after the disturbance amplitude has returned to the normal range is used as input. The autoregressive moving average model is used to fit this sequence, capturing the dynamic changes in load under stable operating conditions. After the model fitting is complete, the load values ​​for subsequent time periods are predicted, and the predicted values ​​are compared with the actual monitored values ​​to calculate the prediction residuals. To quantitatively determine whether the residual sequence is white noise, the Ljung-Box test is used for statistical verification: the test statistic and its corresponding p-value are calculated for the residual sequence at a specified lag order. If the p-value is greater than the significance level, the hypothesis that the residual is white noise is accepted, indicating that the residual fluctuates randomly around zero and has no significant autocorrelation. The current load change conforms to the dynamic characteristics under normal operating conditions of the unit, and the system has returned to stability. If the p-value is less than or equal to the preset level, it indicates that the residual sequence has significant autocorrelation or a systematic deviation from zero. Although the disturbance has disappeared, the system still has potential instability factors and requires further adjustment. Through this stability verification based on model prediction and residual analysis, the recovery status of the thermal power unit after a brief disturbance can be accurately determined, providing a reliable basis for generating subsequent disturbance recovery adjustment commands and avoiding control deviations caused by incomplete system stability. Based on the verification results of the operating status, disturbance recovery adjustment commands are generated. If the verification is successful, a recovery command containing a recovery confirmation flag and a timestamp is generated to mark that the disturbance has ended and to notify the subsequent control module that the system has returned to stability. If the verification fails, an instruction containing an instability warning flag is generated, specifying the time window for continued monitoring. The subsequent control module then decides whether to trigger further adjustments.

[0053] Through the above steps, a complete processing flow for short-term interference type deviations is realized, from interference amplitude monitoring and recovery judgment to stability verification and recovery command generation, ensuring that thermal power units can quickly and reliably recover to a stable operating state after being subjected to external interference.

[0054] Step S6: Integrate stable operation parameter configuration and disturbance recovery adjustment instructions to generate a comprehensive operation control sequence, verify the adaptability of the control sequence under different environmental conditions, and determine the final load adaptive operation adjustment output.

[0055] Step S6 includes: extracting the deviation compensation range and adjustment priority from the stable operation parameter configuration; matching and fusing the recovery timing in the disturbance recovery adjustment command with the deviation compensation range to generate a comprehensive operation control sequence; based on the environmental parameter impact analysis, selecting temperature and humidity as key variables to construct a simulation input set, generating operation scenarios under various external conditions, covering normal, extreme, and transitional environmental conditions; inputting the comprehensive operation control sequence into the operation scenarios for iterative testing, calculating the deviation index under each scenario and comparing it with a preset threshold to verify the adaptability of the control sequence; if the deviation index under all scenarios meets the preset threshold requirements, then determining the final load adaptive operation adjustment output.

[0056] Specifically, after completing the configuration of stable operating parameters and the generation of disturbance recovery adjustment instructions, in order to ensure the reliability and adaptability of the final control strategy under different environmental conditions, the two need to be integrated into a comprehensive operating control sequence, and its performance needs to be verified by simulating a variety of typical operating conditions.

[0057] First, the stable operating parameter configuration is a combination of parameters determined through iterative adjustments to address long-term wear-type deviations. It includes target values, deviation compensation ranges, and adjustment priorities for each operating parameter, such as specific adjustments for parameters like fuel quantity, water flow rate, and air supply volume. The disturbance recovery adjustment command is a recovery command generated after the disturbance subsides, addressing temporary disturbance-type deviations. This command includes not only a recovery confirmation flag and timestamp but also specific parameter recovery operations. For example, it might restore parameters temporarily adjusted due to the disturbance to their pre-disturbance state or execute a set of smooth transition control operations based on the recovered system state. Therefore, the disturbance recovery adjustment command is essentially a sequence of control commands, not just a single flag.

[0058] When merging the two to generate a comprehensive operational control sequence, the adjustment instructions and their priorities for each parameter in the stable operation parameter configuration are extracted first, along with the recovery operation sequence from the disturbance recovery adjustment instructions. The fusion process is aligned along the time axis: recovery operations in the disturbance recovery adjustment instructions typically need to be executed immediately after the disturbance subsides, thus having the highest execution priority; adjustment instructions in the stable operation parameter configuration are executed in their original priority order after the recovery operations are completed. If conflicting control instructions for different parameters occur at the same time during the fusion process—for example, the stable operation parameter configuration requires increasing fuel while the disturbance recovery adjustment instructions require decreasing fuel—the disturbance recovery adjustment instructions take precedence to ensure system stability after the disturbance subsides. If the conflict occurs outside of a recovery period, the parameter instruction with the higher adjustment priority is executed; that is, the parameter instruction ranked earlier in the priority list takes precedence. In this way, the two sets of control logic are integrated into a complete, time-clear comprehensive operational control sequence.

[0059] To verify the adaptability of the integrated operation control sequence under different environmental conditions, multiple operating scenarios need to be constructed in conjunction with environmental parameter impact analysis. Environmental parameter impact analysis refers to analyzing the influence of external temperature and humidity on the load fluctuations and deviation evolution of thermal power units. For example, increased temperature may lead to increased load deviation, and increased humidity may prolong the duration of deviation. Based on this analysis, temperature and humidity are selected as key variables to construct a simulation input set. The correlation characteristics between temperature and load, and the influence of humidity on the duration of deviation are extracted from historical data to generate operating scenarios under various external conditions, including typical operating conditions such as high temperature and high humidity, low temperature and low humidity, extreme temperature fluctuations, and transitional seasons, ensuring that the scenarios cover normal, extreme, and transitional environmental conditions. By simulating load fluctuations under different combinations of environmental parameters, the performance of the control sequence under complex operating conditions can be pre-assessed.

[0060] The integrated operation control sequence is input into various operating scenarios for iterative testing. Iterative testing involves simulating each scenario ten times, and the average value of each deviation index from the ten calculations is taken as the representative value for that scenario to eliminate the influence of random factors. Deviation indices for each scenario are calculated, such as the average and maximum load deviations and the duration of deviations. These calculated deviation indices are compared with preset thresholds. If the deviation indices for all scenarios meet the preset threshold requirements, it indicates that the integrated operation control sequence has good environmental adaptability and can maintain stable and effective control performance under different external conditions. Finally, if the verification is successful, the integrated operation control sequence is determined as the final load adaptive operation adjustment output to guide the actual operation adjustment of thermal power units. This output includes complete control strategies for different deviation types and environmental conditions, enabling adaptive operation adjustment of thermal power units under complex operating conditions. Through the above steps, the fusion generation and environmental adaptability verification of the thermal power unit load adaptive control strategy are achieved, ensuring that the final control output maintains stable and reliable control performance under different environmental conditions.

[0061] The above describes a thermal power unit load adaptive predictive control method according to an embodiment of this application. The following describes a thermal power unit load adaptive predictive control system according to an embodiment of this application. Please refer to [link / reference]. Figure 2 An adaptive load prediction control system for thermal power units, as described in this application embodiment, includes: The data acquisition module is used to acquire the operating load data and environmental parameters of the thermal power unit, analyze the time-series variation pattern of the load deviation, and determine the characteristic description of the deviation. The judgment module is used to analyze the correlation between load deviation and equipment operating status based on the deviation characteristic description, and to judge the nature classification result of the load deviation, which includes long-term wear type and short-term interference type. The prediction module is used to assess the impact of load deviation on unit efficiency for the long-term wear type, determine the wear impact quantification result, predict the future load fluctuation trend and load deviation evolution path of the thermal power unit based on the wear impact quantification result, and determine the preliminary adjustment parameter set. The adjustment module is used to dynamically adjust the operating parameters of the thermal power unit based on the preliminary adjustment parameter set, determine whether the adjusted load deviation meets the preset deviation threshold, and determine the stable operating parameter configuration. The verification module is used to continuously monitor the changes in the interference amplitude for the transient interference type, and verify the system operation status after the interference amplitude recovers to the normal range, and generate interference recovery adjustment instructions based on the verification results. The configuration module is used to integrate the stable operation parameter configuration with the disturbance recovery adjustment command to generate a comprehensive operation control sequence, verify the adaptability of the control sequence under different environmental conditions, and determine the final load adaptive operation adjustment output.

[0062] In summary, this application proposes a method and system for adaptive predictive control of thermal power unit load, constructing a closed-loop adaptive control framework encompassing data acquisition, feature extraction, deviation classification, impact quantification, parameter prediction, dynamic adjustment, disturbance recovery, and environmental adaptability verification. By real-time acquisition of operating load and environmental parameters, the deviation sequence is corrected using environmental parameters, eliminating environmental interference and providing an accurate data foundation for subsequent diagnosis. A two-layer diagnostic mechanism combining historical waveform comparison and time-series feature judgment is employed to accurately distinguish between long-term wear and short-term disturbances. For long-term wear, efficiency loss is quantified through energy balance analysis, and the deviation evolution path is predicted using Monte Carlo simulation. PID control is used to calculate compensation values ​​and decompose them into a multi-parameter adjustment set, dynamically adjusting according to response time priority and achieving closed-loop iterative convergence. For short-term disturbances, recovery commands are generated through disturbance amplitude monitoring and stability verification. Finally, the two control strategies are integrated to generate a comprehensive operating control sequence, and its adaptability is verified through multiple environmental scenarios. This application significantly improves the accuracy, adaptability, and operational stability of thermal power unit load control.

[0063] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0064] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0065] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for adaptive predictive control of thermal power unit load, characterized in that, The method includes: Step S1: Obtain the operating load data and environmental parameters of the thermal power unit, analyze the temporal variation law of the load deviation, and determine the deviation characteristic description; Step S2: Based on the deviation characteristic description, analyze the correlation between load deviation and equipment operating status, and determine the nature classification result of load deviation. The nature classification result includes long-term wear type and short-term disturbance type. Step S3: For the aforementioned long-term wear type, assess the impact of load deviation on unit efficiency, determine the quantification results of wear impact, and based on the quantification results of wear impact, predict the future load fluctuation trend and load deviation evolution path of the thermal power unit, and determine the preliminary set of adjustment parameters. Step S4: Based on the preliminary adjustment parameter set, dynamically adjust the operating parameters of the thermal power unit, determine whether the adjusted load deviation meets the preset deviation threshold, and determine the stable operating parameter configuration; Step S5: For the aforementioned transient interference type, continuously monitor the changes in interference amplitude, and verify the system operating status after the interference amplitude recovers to the normal range, and generate interference recovery adjustment instructions based on the verification results; Step S6: Integrate the stable operation parameter configuration with the disturbance recovery adjustment command to generate a comprehensive operation control sequence, verify the adaptability of the control sequence under different environmental conditions, and determine the final load adaptive operation adjustment output.

2. The adaptive predictive control method for thermal power unit load according to claim 1, characterized in that, Step S1 includes: The operating load data of thermal power units is acquired through a real-time data acquisition system, and environmental parameters such as temperature and humidity are collected simultaneously. Calculate the load change sequence based on the operating load data, and calculate the difference between the actual load and the expected load as the deviation value sequence; The environmental parameters are used to correct the deviation value sequence, eliminating the deviation fluctuation component caused by environmental factors, to obtain the corrected deviation value sequence. Based on the corrected deviation value sequence, the duration and magnitude of the deviation are determined, and a deviation feature description is generated.

3. The adaptive predictive control method for thermal power unit load according to claim 1, characterized in that, Step S2 includes: The time-series waveform data of the current load deviation is obtained, and the time-series waveform data of the current load deviation is compared with the historical deviation waveform data in the historical database. The waveform similarity is calculated by using a dynamic time warping algorithm, wherein the historical deviation waveform data is associated with the corresponding equipment status record. If the waveform similarity exceeds a preset threshold, the equipment status change corresponding to the current load deviation is determined based on the associated equipment status record, and the corresponding property classification result is located based on the equipment status change. Otherwise, if the duration of the deviation in the deviation feature description exceeds a preset threshold and the amplitude change trend shows a gradual shift, then the nature classification result of the load deviation is determined to be long-term wear type. If the duration of the deviation does not exceed the preset threshold, the nature of the load deviation is classified as a transient disturbance.

4. The adaptive predictive control method for thermal power unit load according to claim 1, characterized in that, Step S3 assesses the impact of load deviation on unit efficiency and determines the quantitative results of wear impact, including: For the long-term wear type in the aforementioned property classification results, historical equipment operating status data is obtained. Based on the operating status data, an energy consumption balance analysis method is used to calculate the impact index of load deviation on unit efficiency. The difference between the impact index and the standard efficiency benchmark is calculated, and the difference is corrected in conjunction with environmental parameters. The specific degree of impact of the deviation on the overall operating efficiency of the unit is evaluated, and a quantitative result of wear impact is generated.

5. The adaptive predictive control method for thermal power unit load according to claim 4, characterized in that, Using energy balance analysis, the following indicators are used to calculate the impact of load deviation on unit efficiency: The system acquires input and output energy data from historical equipment operating status data, calculates energy loss by comparing the balance between input and output energy, and divides the energy loss by the total input energy to obtain the efficiency loss rate, which serves as an indicator of the impact of load deviation on unit efficiency.

6. The adaptive predictive control method for thermal power unit load according to claim 1, characterized in that, Step S3 determines the set of adjustment parameters, including: Based on the quantification results of the wear impact, time series analysis is used to analyze the future load fluctuation trend and generate a load fluctuation trend curve. Key points are extracted from the load fluctuation trend curve, and Monte Carlo simulation is used in conjunction with environmental parameters to simulate the deviation evolution path under various environmental scenarios, and the possible range of deviation change under each path is determined. Based on the possible range of the deviation change, a targeted deviation compensation value is calculated using proportional-integral-derivative control logic, and the deviation compensation values ​​are summarized to generate a preliminary set of adjustment parameters.

7. The adaptive predictive control method for thermal power unit load according to claim 1, characterized in that, Step S4 includes: Extract the deviation compensation range and corresponding response time of each parameter from the initial set of adjustment parameters, sort the adjustment parameters according to the response time, and generate an adjustment priority list; According to the aforementioned adjustment priority list, the operating parameters are dynamically adjusted first for the parts with higher deviation severity, and the changes in load deviation value are continuously monitored through a real-time data acquisition system after adjustment. If the detected load deviation value is lower than the preset threshold, the current operating parameter configuration will be determined as the stable operating parameter configuration. If the monitored load deviation value is not lower than the preset threshold, historical equipment operating status data is reacquired, the wear impact quantification results are updated, and the adjustment parameters are recalculated and iteratively adjusted until the load deviation value is lower than the preset threshold.

8. The adaptive predictive control method for thermal power unit load according to claim 1, characterized in that, Step S5 includes: For the transient interference type in the aforementioned property classification results, the current interference amplitude is calculated through real-time load fluctuation monitoring, and the changing trend of the interference amplitude is analyzed. Determine whether the interference amplitude has recovered to the preset normal range. If it has recovered, use time series analysis to verify the stability of the recovered load data and verify the system operating status. Based on the verification results of the operating status, an interference recovery adjustment command is generated.

9. The adaptive predictive control method for thermal power unit load according to claim 1, characterized in that, Step S6 includes: Extract the deviation compensation range and adjustment priority from the stable operation parameter configuration, match and fuse the recovery timing in the interference recovery adjustment command with the deviation compensation range, and generate a comprehensive operation control sequence; Based on the environmental parameter impact analysis, temperature and humidity are selected as key variables to construct a simulation input set, generating operating scenarios under various external conditions. These operating scenarios cover normal, extreme, and transitional environmental conditions. The integrated operation control sequence is input into the operation scenario for iterative testing. The deviation index under each scenario is calculated and compared with a preset threshold to verify the adaptability of the control sequence. If the deviation indicators in all scenarios meet the preset threshold requirements, then the final load adaptive operation adjustment output is determined.

10. A thermal power unit load adaptive predictive control system, used to implement the thermal power unit load adaptive predictive control method as described in any one of claims 1-9, characterized in that, The system includes: The data acquisition module is used to acquire the operating load data and environmental parameters of the thermal power unit, analyze the time-series variation pattern of the load deviation, and determine the characteristic description of the deviation. The judgment module is used to analyze the correlation between load deviation and equipment operating status based on the deviation characteristic description, and to judge the nature classification result of the load deviation, which includes long-term wear type and short-term interference type. The prediction module is used to assess the impact of load deviation on unit efficiency for the long-term wear type, determine the wear impact quantification result, predict the future load fluctuation trend and load deviation evolution path of the thermal power unit based on the wear impact quantification result, and determine the preliminary adjustment parameter set. The adjustment module is used to dynamically adjust the operating parameters of the thermal power unit based on the preliminary adjustment parameter set, determine whether the adjusted load deviation meets the preset deviation threshold, and determine the stable operating parameter configuration. The verification module is used to continuously monitor the changes in the interference amplitude for the transient interference type, and verify the system operation status after the interference amplitude recovers to the normal range, and generate interference recovery adjustment instructions based on the verification results. The configuration module is used to integrate the stable operation parameter configuration with the disturbance recovery adjustment command to generate a comprehensive operation control sequence, verify the adaptability of the control sequence under different environmental conditions, and determine the final load adaptive operation adjustment output.

Images